[EVM] Fold identical BBs after call to jumps lowering

[akiramenai]

In EVM necessity to mark jump destinations with JUMPDEST makes call a terminator, and after its expansion to

	PUSH4 @.RET_DEST
	PUSH4 @func
	JUMP
.RET_DEST:
	JUMPDEST
        <original BB remainder>

Identical basic blocks might appear.
For instance:

declare void @foo()
declare void @bar()
; CHECK-LABEL: diamond
define void @diamond(i1 %par) nounwind {
entry:
  br i1 %par, label %bb1, label %bb2
bb1:
  call void @foo()
  br label %exit
bb2:
  call void @bar()
  br label %exit
exit:
  ret void
}

produces

diamond:                                ; @diamond
; %bb.0:                                ; %entry
	JUMPDEST
	PUSH1 1
	AND
	ISZERO
	PUSH4 @.BB0_2
	JUMPI
.BB0_1:                                 ; %bb1
	JUMPDEST
	PUSH4 @.FUNC_RET0
	PUSH4 @foo
	JUMP
.FUNC_RET0:
	JUMPDEST
	JUMP
.BB0_2:                                 ; %bb2
	JUMPDEST
	PUSH4 @.FUNC_RET1
	PUSH4 @bar
	JUMP
.FUNC_RET1:
	JUMPDEST
	JUMP

and FUNC_RET0, FUNC_RET1 can be merged.
Also, consider moving #722 to MIR level.

[vladimirradosavljevic]

For this,BranchFolding, TailDuplication and MachineBlockPlacement optimizations should help, so I don't think we need to move #722 to MIR level.
These optimizations should also help to optimize select function from #717 (I'm not sure why that example is in that ticket):

LLVM IR:
define i256 @select(i256 %v1, i256 %v2, i256 %v3, i256 %v4) {
  %1 = icmp ne i256 %v3, %v4
  %2 = select i1 %1, i256 %v1, i256 %v2
  ret i256 %2
}

ASM:
select:                                 ; @select
; %bb.0:
	JUMPDEST
	SWAP2
	SWAP1
	SWAP3
	EQ
	ISZERO
	PUSH4 @.BB0_3
	JUMPI
	PUSH4 @.BB0_1
	JUMP
.BB0_3:
	JUMPDEST
	SWAP2
	SWAP1
	POP
	PUSH4 @.BB0_2
	JUMP
.BB0_1:
	JUMPDEST
	POP
	SWAP1
.BB0_2:
	JUMPDEST
	JUMP

to

select:                                 ; @select
; %bb.0:
	JUMPDEST
	SWAP2
	SWAP1
	SWAP3
	EQ
	PUSH4 @.BB0_1
	JUMPI
.BB0_3:
	SWAP2
	SWAP1
	POP
	JUMP
.BB0_1:
	JUMPDEST
	POP
	SWAP1
	JUMP